This deliverable documents the development of the CO2 regional inverse modelling systems WRF-CTDAS and CIF-CHIMERE in preparation for the use of satellite retrievals from the CO2M mission for estimating CO2 fluxes from Western European countries. To this end, a pseudo dataset of CO2M satellite retrievals has been made based on an orbit simulator, a CO2 nature run from the Integrated Forecasting System (IFS) model, and a parameterization of random and systematic retrieval uncertainties. This dataset is used in the WRF-CTDAS inversions to test their ability to recover the fossil and biological CO2 fluxes that were used in the nature run. The inversion setup is the same as used in the national scale inversions reported in CoCO2 deliverable D4.6. Therefore, we test the ability of the inversions to correct the priors used in that inversion setup to those used in the nature run, which are different and
reflect the remaining uncertainties in the spatio-temporal distribution of CO2 fluxes across the domain. In addition, we investigate the impact of uneven sampling by the CO2M satellite constellation due to cloud cover by redistributing the retrievals randomly across the orbit. Inversions are performed for winter and summer to test seasonal differences in the CO2M inversion performance and the impact of resampling. The results show that resampling has a significant effect. However, the inversion setups have not reached the state yet that robust conclusions can be derived. The deliverables also documents plans for inversions with the CIF-CHIMERE system propagating the retrieval errors from the pseudo CO2M observations, but also of pseudo OCO-2 and surface observations, to quantify the uncertainty in the corresponding emission estimates and demonstrate the added value of the CO2M mission.
Abstract
Files